InDRiVE: Reward-Free World-Model Pretraining for Autonomous Driving via Latent Disagreement

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Khanzada, Feeza Khan, Kwon, Jaerock
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917171406307328
author Khanzada, Feeza Khan
Kwon, Jaerock
author_facet Khanzada, Feeza Khan
Kwon, Jaerock
contents Model-based reinforcement learning (MBRL) can reduce interaction cost for autonomous driving by learning a predictive world model, but it typically still depends on task-specific rewards that are difficult to design and often brittle under distribution shift. This paper presents InDRiVE, a DreamerV3-style MBRL agent that performs reward-free pretraining in CARLA using only intrinsic motivation derived from latent ensemble disagreement. Disagreement acts as a proxy for epistemic uncertainty and drives the agent toward under-explored driving situations, while an imagination-based actor-critic learns a planner-free exploration policy directly from the learned world model. After intrinsic pretraining, we evaluate zero-shot transfer by freezing all parameters and deploying the pretrained exploration policy in unseen towns and routes. We then study few-shot adaptation by training a task policy with limited extrinsic feedback for downstream objectives (lane following and collision avoidance). Experiments in CARLA across towns, routes, and traffic densities show that disagreement-based pretraining yields stronger zero-shot robustness and robust few-shot collision avoidance under town shift and matched interaction budgets, supporting the use of intrinsic disagreement as a practical reward-free pretraining signal for reusable driving world models.
format Preprint
id arxiv_https___arxiv_org_abs_2512_18850
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle InDRiVE: Reward-Free World-Model Pretraining for Autonomous Driving via Latent Disagreement
Khanzada, Feeza Khan
Kwon, Jaerock
Robotics
Model-based reinforcement learning (MBRL) can reduce interaction cost for autonomous driving by learning a predictive world model, but it typically still depends on task-specific rewards that are difficult to design and often brittle under distribution shift. This paper presents InDRiVE, a DreamerV3-style MBRL agent that performs reward-free pretraining in CARLA using only intrinsic motivation derived from latent ensemble disagreement. Disagreement acts as a proxy for epistemic uncertainty and drives the agent toward under-explored driving situations, while an imagination-based actor-critic learns a planner-free exploration policy directly from the learned world model. After intrinsic pretraining, we evaluate zero-shot transfer by freezing all parameters and deploying the pretrained exploration policy in unseen towns and routes. We then study few-shot adaptation by training a task policy with limited extrinsic feedback for downstream objectives (lane following and collision avoidance). Experiments in CARLA across towns, routes, and traffic densities show that disagreement-based pretraining yields stronger zero-shot robustness and robust few-shot collision avoidance under town shift and matched interaction budgets, supporting the use of intrinsic disagreement as a practical reward-free pretraining signal for reusable driving world models.
title InDRiVE: Reward-Free World-Model Pretraining for Autonomous Driving via Latent Disagreement
topic Robotics
url https://arxiv.org/abs/2512.18850